Figure 1 created from figure1.do and gasprice.txt:
figure1.do reads in price indexes (from gasprice.txt) for gasoline (the chain-price index for gasoline and other motor fuel published by the 
Bureau of Labor Statistics) and for all other consumer goods and services (calculate by the authors as the result from chain-subtracting the 
gasoline price index from the chain-price index for all personal consumption expenditures).  Both indexes equal 100 in 2005, so the result is 
multiplied by the average price of gasoline in 2005 as reported by the US Energy Information Administration (all grades, areas and formulations, 
reported in column B of the "Data 12" spreadsheet downloaded from the "full history" link at http://www.eia.gov/petroleum/gasdiesel/).  

Unit root tests for real gasoline prices: 
RATS program is gaspriceann.prg, which uses data from gaspriceann.txt and the procedure DFUNIT.SRC.  Output is gaspriceann.out

Figure 2:
The West Texas Intermediate spot price is from the Commodity Research Bureau
NYMEX Crude Oil futures are from the US Energy Information Administration.  Data for each month are the 2-year ahead futures price for that month.
Data are reported in figure2.txt.

Table 1:
House prices sample is zip19802009_panel.dta.  House price data are proprietary (purchased from CoreLogic) and have been dropped.  Other variables are:
count = number of ZIP codes per MSA
landarea1990 = land area in 1990
units1980 = number of housing units in 1980
m2_1980 = average commute time in 1980
Dataset was created by gphv_setup.do and log file showing summary stats is gphv_setup.log
Construction sample is placeann.dta.  Variables are:
nplace2 = number of places in MSA
areal90 = place land area in 1990
hu80 = number of housing units in 1980
avgtime80 = average commute time in 1980
the sample used in the paper is marked with the variable cdata = 1
rgasp = the annual average of the chain-price index for gasoline divided by the annual average of the chain-price index for other consumer goods
and services mulitplied by the average price of gasoline in 2005.
dlrgasp = change in ln(rgasp). Mean, median and standard deviation of the percent change reported in the table are for 1977 to 2007, since those are
the years used in the main analysis.
Summary statistics for the construction sample reported in Table 1 (except the change in the real gas price) are shown in placeann_stats.do

Table 2:
House price results are based on proprietary data.
Construction results in table2.do, pulling from placeann.dta
Note that there is a typographical error in the 7th row of the table as published.  The sum of coefficients on lagged GDP growth times commute time 
should be -0.058 (as shown in the log file) not 0.058.

Figure 3:
figure3.do, pulling from placeann.dta

Figure 4:
Results are based on proprietary data.

Table 3:
House price results are based on proprietary data.
Construction results in table3.do, pulling from placeann.dta


Table 4:
House price results are based on proprietary data.
Construction results in table4.do, pulling from placeann.dta and mfgshare.dta
mfgshare.dta has the share of manufacturing employment by metropolitan area from the 1980 Census, created in mfgshare.do
wharton.dta has the Wharton regulation data, created in wharton.do

